Help:How do I search FactGrid?

From FactGrid
Jump to: navigation, search


<provisional needs revision>

Basic Item and full text search

The search field at the top right offers two different search functions:

  • In bold, hits in the database are displayed first. Alias versions are included here.
  • The last field of the autocomplete menue offers a focused text search - a text search looking at labels, descriptions and independent pages (excluding separate name spaces such as discussion pages.

If you want to run a full text search of all the name spaces use this link from the menue, you can here select different FactGrid sections like help pages or Item discussion pages or project pages:

What links here

The menu's What links here link search is one on the most useful links in the entire menue. Use it if you have landed on a page with totally insufficient information in order to understand why this page has been created. You will get a list of all the pages that refer to the page you are on and you can use this knowledge to gain more information about this item.

The query is also needed after you have changed a title or merged two Items to get a list of the old refereces which you might like to manually reset to the new address.

The more complex searches are compiled in SPARQL

FactGrid can be queried in any imaginable complexity and the results can be displayed in tables, network graphs, on timelines or other visualisation. These more complex searches are run at the SPARQL endpoint:

There are plenty of introductions to SPARQL on the Internet. In practice, the query language is a significant hurdle; even the first steps are not clear: there is an input help that but that has to be opened with the "i" icon. If you click the icon, the two part screen opens up into a three part division: you will now be able to see how the interface writes SPARQL code for you - basic code for basic searches. More complex searches require direct scripting, which is not that trivial. The two central Wikidata manuals are:

Both links are of limited use on FactGrid, since we are not sharing the data models (every new Wikibase installation is counting P-Numbers its own way and labels and design features will differ).

In the simplest search query you use the first input to "filter": You define which database objects you want to access with your search, for instance all "humans" - the input is here P2—Q7 you do not have to know the numbers, the query service is offering you the databases auto complete suggestions.

In the second step you should decide what exactly you want to see — with a command that can now ask for all the dates of birth for instance.

Press the blue button with the arrow (triangle) to activate the search.

In order to understand what you can search for you need to have an idea of the properties on FactGrid — they create the statements which you might like to comb through.

With all these deficits, SPARQL remains a brilliant search language, since it as open as the statements you can make complex enough to go into the depth of data connections. In SPARQL queries and visualizations are quickly possible.

"Sample queries" are the best practical introduction to SPARQL

In practice, "sample queries" prove to be the first choice to understand different ways to search the database. With them you receive practical searches, which you can now modify to suit your own interest. Once on SPARQL you should again open the three-window-scree by pressing the i-icon. You can now use the left upper window to odify the search and the right upper window to study the script.

See the query in your language

SPARQL queries are usually conducted in your browser language. If the query is not performed in the language you want to see in the results use the language setting on top/right.

The future: independent user interfaces

The future of Wikibase will foreseeably lie in a plurality of interfaces created by projects who use the database. The interfaces will offer search fields and do the SPARQL to communicate with the database.

See our blog's article for a look at such solutions tested with the Wikidata database:

  • "Archive guide to the German colonial period" online - a conversation with Uwe Jung, Potsdam University of Applied Sciences, about the use of Wikidata as a research platform